- djm@cvs.openbsd.org 2010/09/25 09:30:16

     [sftp.c configure.ac openbsd-compat/glob.c openbsd-compat/glob.h]
     make use of new glob(3) GLOB_KEEPSTAT extension to save extra server
     rountrips to fetch per-file stat(2) information.
     NB. update openbsd-compat/ glob(3) implementation from OpenBSD libc to
     match.

1 files changed, 170 insertions, 62 deletions

diff --git a/openbsd-compat/glob.c b/openbsd-compat/glob.c
index 74b50640..7bbe6c71 100644
--- a/openbsd-compat/glob.c
+++ b/openbsd-compat/glob.c
@@ -1,4 +1,4 @@
-/*	$OpenBSD: glob.c,v 1.26 2005/11/28 17:50:12 deraadt Exp $ */
+/*	$OpenBSD: glob.c,v 1.33 2010/09/26 22:15:39 djm Exp $ */
 /*
  * Copyright (c) 1989, 1993
  *	The Regents of the University of California.  All rights reserved.
@@ -33,6 +33,31 @@
 
 /* OPENBSD ORIGINAL: lib/libc/gen/glob.c */
 
+/*
+ * glob(3) -- a superset of the one defined in POSIX 1003.2.
+ *
+ * The [!...] convention to negate a range is supported (SysV, Posix, ksh).
+ *
+ * Optional extra services, controlled by flags not defined by POSIX:
+ *
+ * GLOB_QUOTE:
+ *	Escaping convention: \ inhibits any special meaning the following
+ *	character might have (except \ at end of string is retained).
+ * GLOB_MAGCHAR:
+ *	Set in gl_flags if pattern contained a globbing character.
+ * GLOB_NOMAGIC:
+ *	Same as GLOB_NOCHECK, but it will only append pattern if it did
+ *	not contain any magic characters.  [Used in csh style globbing]
+ * GLOB_ALTDIRFUNC:
+ *	Use alternately specified directory access functions.
+ * GLOB_TILDE:
+ *	expand ~user/foo to the /home/dir/of/user/foo
+ * GLOB_BRACE:
+ *	expand {1,2}{a,b} to 1a 1b 2a 2b
+ * gl_matchc:
+ *	Number of matches in the current invocation of glob.
+ */
+
 #include "includes.h"
 
 #include <sys/types.h>
@@ -47,7 +72,7 @@
 #include <unistd.h>
 
 #if !defined(HAVE_GLOB) || !defined(GLOB_HAS_ALTDIRFUNC) || \
-    !defined(GLOB_HAS_GL_MATCHC) || \
+    !defined(GLOB_HAS_GL_MATCHC) || !defined(GLOB_HAS_GL_STATV) || \
     !defined(HAVE_DECL_GLOB_NOMATCH) || HAVE_DECL_GLOB_NOMATCH == 0 || \
     defined(BROKEN_GLOB)
 
@@ -63,31 +88,7 @@ get_arg_max(void)
 #endif
 }
 
-/*
- * glob(3) -- a superset of the one defined in POSIX 1003.2.
- *
- * The [!...] convention to negate a range is supported (SysV, Posix, ksh).
- *
- * Optional extra services, controlled by flags not defined by POSIX:
- *
- * GLOB_QUOTE:
- *	Escaping convention: \ inhibits any special meaning the following
- *	character might have (except \ at end of string is retained).
- * GLOB_MAGCHAR:
- *	Set in gl_flags if pattern contained a globbing character.
- * GLOB_NOMAGIC:
- *	Same as GLOB_NOCHECK, but it will only append pattern if it did
- *	not contain any magic characters.  [Used in csh style globbing]
- * GLOB_ALTDIRFUNC:
- *	Use alternately specified directory access functions.
- * GLOB_TILDE:
- *	expand ~user/foo to the /home/dir/of/user/foo
- * GLOB_BRACE:
- *	expand {1,2}{a,b} to 1a 1b 2a 2b
- * gl_matchc:
- *	Number of matches in the current invocation of glob.
- */
-
+#include "charclass.h"
 
 #define	DOLLAR		'$'
 #define	DOT		'.'
@@ -100,7 +101,6 @@ get_arg_max(void)
 #define	RBRACKET	']'
 #define	SEP		'/'
 #define	STAR		'*'
-#undef TILDE			/* Some platforms may already define it */
 #define	TILDE		'~'
 #define	UNDERSCORE	'_'
 #define	LBRACE		'{'
@@ -137,6 +137,7 @@ typedef char Char;
 #define	M_ONE		META('?')
 #define	M_RNG		META('-')
 #define	M_SET		META('[')
+#define	M_CLASS		META(':')
 #define	ismeta(c)	(((c)&M_QUOTE) != 0)
 
 
@@ -144,7 +145,8 @@ static int	 compare(const void *, const void *);
 static int	 g_Ctoc(const Char *, char *, u_int);
 static int	 g_lstat(Char *, struct stat *, glob_t *);
 static DIR	*g_opendir(Char *, glob_t *);
-static Char	*g_strchr(Char *, int);
+static Char	*g_strchr(const Char *, int);
+static int	 g_strncmp(const Char *, const char *, size_t);
 static int	 g_stat(Char *, struct stat *, glob_t *);
 static int	 glob0(const Char *, glob_t *);
 static int	 glob1(Char *, Char *, glob_t *, size_t *);
@@ -152,11 +154,11 @@ static int	 glob2(Char *, Char *, Char *, Char *, Char *, Char *,
 		    glob_t *, size_t *);
 static int	 glob3(Char *, Char *, Char *, Char *, Char *,
 		    Char *, Char *, glob_t *, size_t *);
-static int	 globextend(const Char *, glob_t *, size_t *);
+static int	 globextend(const Char *, glob_t *, size_t *, struct stat *);
 static const Char *
 		 globtilde(const Char *, Char *, size_t, glob_t *);
 static int	 globexp1(const Char *, glob_t *);
-static int	 globexp2(const Char *, const Char *, glob_t *, int *);
+static int	 globexp2(const Char *, const Char *, glob_t *);
 static int	 match(Char *, Char *, Char *);
 #ifdef DEBUG
 static void	 qprintf(const char *, Char *);
@@ -174,6 +176,7 @@ glob(const char *pattern, int flags, int (*errfunc)(const char *, int),
 	if (!(flags & GLOB_APPEND)) {
 		pglob->gl_pathc = 0;
 		pglob->gl_pathv = NULL;
+		pglob->gl_statv = NULL;
 		if (!(flags & GLOB_DOOFFS))
 			pglob->gl_offs = 0;
 	}
@@ -215,15 +218,13 @@ static int
 globexp1(const Char *pattern, glob_t *pglob)
 {
 	const Char* ptr = pattern;
-	int rv;
 
 	/* Protect a single {}, for find(1), like csh */
 	if (pattern[0] == LBRACE && pattern[1] == RBRACE && pattern[2] == EOS)
 		return glob0(pattern, pglob);
 
-	while ((ptr = (const Char *) g_strchr((Char *) ptr, LBRACE)) != NULL)
-		if (!globexp2(ptr, pattern, pglob, &rv))
-			return rv;
+	if ((ptr = (const Char *) g_strchr(ptr, LBRACE)) != NULL)
+		return globexp2(ptr, pattern, pglob);
 
 	return glob0(pattern, pglob);
 }
@@ -235,9 +236,9 @@ globexp1(const Char *pattern, glob_t *pglob)
  * If it fails then it tries to glob the rest of the pattern and returns.
  */
 static int
-globexp2(const Char *ptr, const Char *pattern, glob_t *pglob, int *rv)
+globexp2(const Char *ptr, const Char *pattern, glob_t *pglob)
 {
-	int     i;
+	int     i, rv;
 	Char   *lm, *ls;
 	const Char *pe, *pm, *pl;
 	Char    patbuf[MAXPATHLEN];
@@ -270,10 +271,8 @@ globexp2(const Char *ptr, const Char *pattern, glob_t *pglob, int *rv)
 		}
 
 	/* Non matching braces; just glob the pattern */
-	if (i != 0 || *pe == EOS) {
-		*rv = glob0(patbuf, pglob);
-		return 0;
-	}
+	if (i != 0 || *pe == EOS)
+		return glob0(patbuf, pglob);
 
 	for (i = 0, pl = pm = ptr; pm <= pe; pm++) {
 		switch (*pm) {
@@ -319,7 +318,9 @@ globexp2(const Char *ptr, const Char *pattern, glob_t *pglob, int *rv)
 #ifdef DEBUG
 				qprintf("globexp2:", patbuf);
 #endif
-				*rv = globexp1(patbuf, pglob);
+				rv = globexp1(patbuf, pglob);
+				if (rv && rv != GLOB_NOMATCH)
+					return rv;
 
 				/* move after the comma, to the next string */
 				pl = pm + 1;
@@ -330,7 +331,6 @@ globexp2(const Char *ptr, const Char *pattern, glob_t *pglob, int *rv)
 			break;
 		}
 	}
-	*rv = 0;
 	return 0;
 }
 
@@ -399,6 +399,47 @@ globtilde(const Char *pattern, Char *patbuf, size_t patbuf_len, glob_t *pglob)
 	return patbuf;
 }
 
+static int
+g_strncmp(const Char *s1, const char *s2, size_t n)
+{
+	int rv = 0;
+
+	while (n--) {
+		rv = *(Char *)s1 - *(const unsigned char *)s2++;
+		if (rv)
+			break;
+		if (*s1++ == '\0')
+			break;
+	}
+	return rv;
+}
+
+static int
+g_charclass(const Char **patternp, Char **bufnextp)
+{
+	const Char *pattern = *patternp + 1;
+	Char *bufnext = *bufnextp;
+	const Char *colon;
+	struct cclass *cc;
+	size_t len;
+
+	if ((colon = g_strchr(pattern, ':')) == NULL || colon[1] != ']')
+		return 1;	/* not a character class */
+
+	len = (size_t)(colon - pattern);
+	for (cc = cclasses; cc->name != NULL; cc++) {
+		if (!g_strncmp(pattern, cc->name, len) && cc->name[len] == '\0')
+			break;
+	}
+	if (cc->name == NULL)
+		return -1;	/* invalid character class */
+	*bufnext++ = M_CLASS;
+	*bufnext++ = (Char)(cc - &cclasses[0]);
+	*bufnextp = bufnext;
+	*patternp += len + 3;
+
+	return 0;
+}
 
 /*
  * The main glob() routine: compiles the pattern (optionally processing
@@ -427,7 +468,7 @@ glob0(const Char *pattern, glob_t *pglob)
 			if (c == NOT)
 				++qpatnext;
 			if (*qpatnext == EOS ||
-			    g_strchr((Char *) qpatnext+1, RBRACKET) == NULL) {
+			    g_strchr(qpatnext+1, RBRACKET) == NULL) {
 				*bufnext++ = LBRACKET;
 				if (c == NOT)
 					--qpatnext;
@@ -438,6 +479,20 @@ glob0(const Char *pattern, glob_t *pglob)
 				*bufnext++ = M_NOT;
 			c = *qpatnext++;
 			do {
+				if (c == LBRACKET && *qpatnext == ':') {
+					do {
+						err = g_charclass(&qpatnext,
+						    &bufnext);
+						if (err)
+							break;
+						c = *qpatnext++;
+					} while (c == LBRACKET && *qpatnext == ':');
+					if (err == -1 &&
+					    !(pglob->gl_flags & GLOB_NOCHECK))
+						return GLOB_NOMATCH;
+					if (c == RBRACKET)
+						break;
+				}
 				*bufnext++ = CHAR(c);
 				if (*qpatnext == RANGE &&
 				    (c = qpatnext[1]) != RBRACKET) {
@@ -484,7 +539,7 @@ glob0(const Char *pattern, glob_t *pglob)
 		if ((pglob->gl_flags & GLOB_NOCHECK) ||
 		    ((pglob->gl_flags & GLOB_NOMAGIC) &&
 		    !(pglob->gl_flags & GLOB_MAGCHAR)))
-			return(globextend(pattern, pglob, &limit));
+			return(globextend(pattern, pglob, &limit, NULL));
 		else
 			return(GLOB_NOMATCH);
 	}
@@ -547,7 +602,7 @@ glob2(Char *pathbuf, Char *pathbuf_last, Char *pathend, Char *pathend_last,
 				*pathend = EOS;
 			}
 			++pglob->gl_matchc;
-			return(globextend(pathbuf, pglob, limitp));
+			return(globextend(pathbuf, pglob, limitp, &sb));
 		}
 
 		/* Find end of next segment, copy tentatively to pathend. */
@@ -670,25 +725,40 @@ glob3(Char *pathbuf, Char *pathbuf_last, Char *pathend, Char *pathend_last,
  *	gl_pathv points to (gl_offs + gl_pathc + 1) items.
  */
 static int
-globextend(const Char *path, glob_t *pglob, size_t *limitp)
+globextend(const Char *path, glob_t *pglob, size_t *limitp, struct stat *sb)
 {
 	char **pathv;
-	int i;
-	u_int newsize, len;
-	char *copy;
+	ssize_t i;
+	size_t newn, len;
+	char *copy = NULL;
 	const Char *p;
-
-	newsize = sizeof(*pathv) * (2 + pglob->gl_pathc + pglob->gl_offs);
-	pathv = pglob->gl_pathv ? realloc((char *)pglob->gl_pathv, newsize) :
-	    malloc(newsize);
-	if (pathv == NULL) {
+	struct stat **statv;
+
+	newn = 2 + pglob->gl_pathc + pglob->gl_offs;
+	if (SIZE_MAX / sizeof(*pathv) <= newn ||
+	    SIZE_MAX / sizeof(*statv) <= newn) {
+ nospace:
+		for (i = pglob->gl_offs; i < newn - 2; i++) {
+			if (pglob->gl_pathv && pglob->gl_pathv[i])
+				free(pglob->gl_pathv[i]);
+			if ((pglob->gl_flags & GLOB_KEEPSTAT) != 0 &&
+			    pglob->gl_pathv && pglob->gl_pathv[i])
+				free(pglob->gl_statv[i]);
+		}
 		if (pglob->gl_pathv) {
 			free(pglob->gl_pathv);
 			pglob->gl_pathv = NULL;
 		}
+		if (pglob->gl_statv) {
+			free(pglob->gl_statv);
+			pglob->gl_statv = NULL;
+		}
 		return(GLOB_NOSPACE);
 	}
 
+	pathv = realloc(pglob->gl_pathv, newn * sizeof(*pathv));
+	if (pathv == NULL)
+		goto nospace;
 	if (pglob->gl_pathv == NULL && pglob->gl_offs > 0) {
 		/* first time around -- clear initial gl_offs items */
 		pathv += pglob->gl_offs;
@@ -697,6 +767,29 @@ globextend(const Char *path, glob_t *pglob, size_t *limitp)
 	}
 	pglob->gl_pathv = pathv;
 
+	if ((pglob->gl_flags & GLOB_KEEPSTAT) != 0) {
+		statv = realloc(pglob->gl_statv, newn * sizeof(*statv));
+		if (statv == NULL)
+			goto nospace;
+		if (pglob->gl_statv == NULL && pglob->gl_offs > 0) {
+			/* first time around -- clear initial gl_offs items */
+			statv += pglob->gl_offs;
+			for (i = pglob->gl_offs; --i >= 0; )
+				*--statv = NULL;
+		}
+		pglob->gl_statv = statv;
+		if (sb == NULL)
+			statv[pglob->gl_offs + pglob->gl_pathc] = NULL;
+		else {
+			if ((statv[pglob->gl_offs + pglob->gl_pathc] =
+			    malloc(sizeof(**statv))) == NULL)
+				goto copy_error;
+			memcpy(statv[pglob->gl_offs + pglob->gl_pathc], sb,
+			    sizeof(*sb));
+		}
+		statv[pglob->gl_offs + pglob->gl_pathc + 1] = NULL;
+	}
+
 	for (p = path; *p++;)
 		;
 	len = (size_t)(p - path);
@@ -711,11 +804,11 @@ globextend(const Char *path, glob_t *pglob, size_t *limitp)
 	pathv[pglob->gl_offs + pglob->gl_pathc] = NULL;
 
 	if ((pglob->gl_flags & GLOB_LIMIT) &&
-	    newsize + *limitp >= (u_int) get_arg_max()) {
+	    (newn * sizeof(*pathv)) + *limitp >= ARG_MAX) {
 		errno = 0;
 		return(GLOB_NOSPACE);
 	}
-
+ copy_error:
 	return(copy == NULL ? GLOB_NOSPACE : 0);
 }
 
@@ -751,13 +844,21 @@ match(Char *name, Char *pat, Char *patend)
 				return(0);
 			if ((negate_range = ((*pat & M_MASK) == M_NOT)) != EOS)
 				++pat;
-			while (((c = *pat++) & M_MASK) != M_END)
+			while (((c = *pat++) & M_MASK) != M_END) {
+				if ((c & M_MASK) == M_CLASS) {
+					int idx = *pat & M_MASK;
+					if (idx < NCCLASSES &&
+					    cclasses[idx].isctype(k))
+						ok = 1;
+					++pat;
+				}
 				if ((*pat & M_MASK) == M_RNG) {
 					if (c <= k && k <= pat[1])
 						ok = 1;
 					pat += 2;
 				} else if (c == k)
 					ok = 1;
+			}
 			if (ok == negate_range)
 				return(0);
 			break;
@@ -785,6 +886,14 @@ globfree(glob_t *pglob)
 		free(pglob->gl_pathv);
 		pglob->gl_pathv = NULL;
 	}
+	if (pglob->gl_statv != NULL) {
+		for (i = 0; i < pglob->gl_pathc; i++) {
+			if (pglob->gl_statv[i] != NULL)
+				free(pglob->gl_statv[i]);
+		}
+		free(pglob->gl_statv);
+		pglob->gl_statv = NULL;
+	}
 }
 
 static DIR *
@@ -830,11 +939,11 @@ g_stat(Char *fn, struct stat *sb, glob_t *pglob)
 }
 
 static Char *
-g_strchr(Char *str, int ch)
+g_strchr(const Char *str, int ch)
 {
 	do {
 		if (*str == ch)
-			return (str);
+			return ((Char *)str);
 	} while (*str++);
 	return (NULL);
 }
@@ -870,5 +979,4 @@ qprintf(const char *str, Char *s)
 #endif
 
 #endif /* !defined(HAVE_GLOB) || !defined(GLOB_HAS_ALTDIRFUNC) ||
-          !defined(GLOB_HAS_GL_MATCHC) */
-
+          !defined(GLOB_HAS_GL_MATCHC) || !defined(GLOB_HAS_GL_STATV) */